I completely agree. By and large, the cars live up to the hype and then some, esp with the Model 3 which was designed almost too well that it is cannibalizing S and X sales.
But in terms of Autopilot, it is the exact opposite. I have had AP for 2 years now, and to be honest, the improvement in performance has been marginal. It has tracked the lane and parallel parked very well, but beyond that, I have not seen much improvement at all. I have been using NOA with lane change, and it doesn't work very well IMO. It will change in and out of lanes for no reason sometimes, and it stays in the passing lane far too long, even when a car comes up behind me and starts tailgating. One thing that has improved is that it won't change lanes until it sees the adjacent lane is clear.
It can't even drive very well in a straight line on a city street. It doesn't follow closely enough after the car in front starts moving, and it is jittery around curves. And don't even get me started on those autowipers which still don't work properly after 2 years - and FSD is supposed to be feature complete by the end of this year?
Honestly, I'm guessing that Elon is overhyping Autopilot to promote the cutting edge image of the brand, and to earn extra margin on low margin hardware by making loyal followers "believe". Why else would they sell something that doesn't even exist, and now scare people into buying by raising the price on May 1? And maybe they moved NOA, Summon, and Autopark into FSD because they knew that it will be a long time before real features could be delivered, and this will be a way to offer something in the interim to entice people into buying it.
I hope they can deliver something on Monday, but as another poster has said, the 2016 FSD video was already impressive and nothing has happened since.